Introducing Stereo, OpenAI and the art of changing one’s mind
Ephemeral Reviews, Essays and Opinions s°01.ep01 - 2019.03.25
Dear friends and coworkers,
We are extremely happy to send you the first issue of Stéréo (well, first after our limited #0 release from two weeks ago), our digital-oriented newsletter highlighting the main developments and weak signals affecting the world’s societies and economies.
Once every two weeks, you’ll find the following sections:
Play: Presenting the gist of recent, major news.
Pause: Taking a step back with an editorial.
Forward: Featuring emerging use cases, scientific discoveries and product launches.
Rewind: Unearthing fun facts, figures and stories from the somewhat recent past.
We hope you’ll enjoy reading it as much as we enjoy crafting it.
Stéphane (S) & Tom (T)
Two certainties here. SXSW is the most interesting tech conference year after year and Margrethe Vestager would be a really great president to bring the European Commission into the 21st century. (S) Link
Yet another report on competition at the digital age… but one with clear statements and proposals that could be useful to Europe, issued from an independent panel of international experts gathered by the British Government. UK, please baby don’t go. (S) Link
Fascinating to see how any useful app ends up being a message app. As much as how kids can trick their teachers and parents. (S) Link
A detailed account by a former temp contractor of what it’s like to work inside an Amazon warehouse - not surprisingly, it’s appalling. (T) Link
Only fools don’t change their minds, as the saying goes. Well, now it seems like OpenAI (the research structure founded by Elon Musk and Sam Altman, with $1bn pledged at the outset, and whose mission is “to ensure that artificial general intelligence benefits all of humanity”) didn’t want to keep fooling itself any longer.
OpenAI was born as a non-profit in December 2015, “unconstrained by a need to generate financial return”. So the recent announcement that most of its activities - i.e. research - had been transferred to a for-profit company was slightly surprising. Their rationale: “AI research is really^10 costly (brain & computing power) => we need much more capital => we need to offer long-term returns to potential investors through a for-profit scheme. But don’t worry, we’ll cap returns at 100x (editor’s note: !!!) and our mission will trump financial interests, since the non-profit will control the for-profit.”
(Btw: funny to see how the strict separation between management and capital, one of the oldest company structures, resurfaces within a cutting-edge field.)
To me, it sounds above all like a half-acknowledgement that their grand mission was doomed from the outset (the other proof being Elon Musk’s distanciation from OpenAI one year ago):
AI is a fluid field: ideas, talents, algorithms move around, and may be used with good and bad intents alike. The only “hard” asset that exists is data - and OpenAI has no huge proprietary datasets. Hence pretending that a single group may steer the whole field in a radically new direction is a wonderful story to tell - and nothing more.
As lyrical as their mission may be, their (top-level) research agenda is no different from what fundamental AI research teams are tackling at Google, Microsoft, Facebook…
… so they can’t entice AI researchers through a higher sense of purpose…
… so all things being equal, why would those accept a lower compensation in this burning-hot field?
OpenAI adapting its organization to the reality behind its flamboyant-but-ineffective mission is not a U-turn: it’s a sign of maturity. For sure, “just” being a fundamental AI research company won’t sound as glorious as saving all of humanity. But OpenAI is one of the best that is, and this just might be what the world needs right now.
Google is trying hard to catch up on both cloud infrastructure and the gaming industry. It raises many questions on the future of gaming but access, inclusivity and maybe even fun may get lost in this attempt. (S) Link
Education could morph into very addictive games, and math is a first proof of concept, as shown by this analysis of the Prodigy phenomenon. (T) Link
Legalize it. But with style. Smoke dope like a Rothschild or a Rockefeller. (S) Link
Should memes become copyrightable? (T) Link
Do you know the link between a dead cow and the next president of the United States ? Hacktivism. And you should start following Beto O’Rourke. (S) Link
Can you believe it? China has produced more cement in the 5 last years than the United States in the entire century. (S)
Art is certainly one of the most sophisticated, global and meaningful markets. Many insights and trends to watch in last year transactions. (S) Link
Rating world leaders in 130 countries: 31% approve US leadership… and 34% China’s. (T) Link
A few minutes with a pure object of analogic perfection. The kind of item that was not possible to craft earlier and won’t be possible to make anymore. (S) Link
If you enjoyed this newsletter
Copyright © 2019 FABERNOVEL, All rights reserved.
Edited by Stéphane Distinguin & Tom Morisse
FABERNOVEL is a talent company that creates digital products and services to support companies in their transformation and innovation trajectory.
Congratulations! You've reached the bottom of the page. Here's your reward.